Genome Association Analysis Under Different Data Structures

Due to genetic background and environmental e?ect on population allele fre-quencies, resulting in the phenomenon of population stratification, a limitation ofAssociation analysis is bias can result from population stratification. Now, there aretwo methods to correct this bias: family-based association analysis and population-based case-control test. Family-based association analysis uses data of biologicalrelatives as control. Although to some extent, the phenomenon of population strati-fication can be corrected. It is di?cult to collect data of a?ected individual families.The traditional case-control studies may be subject to bias caused by populationstratification, alternative methods that robust to population stratification such asGenomic control and Structured association, et al. Recently, Qizhai Li et al andShuanglin Zhang et al proposed some two-stage designs. In this article, we proposea similarity-based association analysis using a series of independent markers , thefirst stage is a clustering test that uses similarity defined in Fuzzy mathematics todetermine whether a pair of individuals belong to the same subpopulation, The sec-ond stage is an similarity-based association analysis. Our simulation results showthat the approaches we proposed have control type I error rates and much morepowerful.